Correct expectations and fix H&S in scale means

src/cr/cube/crunch_cube.py

@@ -1128,14 +1128,20 @@ def scale_means(self, hs_dims=None, prune=False):
         slices_means = [ScaleMeans(slice_).data for slice_ in self.slices]
 
         if  hs_dims and self.ndim > 1:
-            # Intersperse scale means with nans if H&S specified
-            inserted_indices = self.inserted_hs_indices()
+            # Intersperse scale means with nans if H&S specified, and 2D. No
+            # need to modify 1D, as only one mean will ever be inserted.
+            inserted_indices = self.inserted_hs_indices()[-2:]
             for scale_means in slices_means:
-                if scale_means[0] is not None and 0 in hs_dims and inserted_indices[0]:
-                    for i in inserted_indices[0]:
-                        scale_means[0] = np.insert(scale_means[0], i, np.nan)
-                if scale_means[1] is not None and 1 in hs_dims and inserted_indices[1]:
+                # Scale means 0 corresonds to the column dimension (is
+                # calculated by using its values). The result of it, however,
+                # is a row. That's why we need to check the insertions on the
+                # row dim (inserted columns).
+                if scale_means[0] is not None and 1 in hs_dims and inserted_indices[1]:
                     for i in inserted_indices[1]:
+                        scale_means[0] = np.insert(scale_means[0], i, np.nan)
+                # Scale means 1 is a column, so we need to check for row insertions.
+                if scale_means[1] is not None and 0 in hs_dims and inserted_indices[0]:
+                    for i in inserted_indices[0]:
                         scale_means[1] = np.insert(scale_means[1], i, np.nan)
 
         if prune:

tests/integration/test_scale_means.py

@@ -136,8 +136,8 @@ def test_cat_x_cat_with_hs():
 
     # Test with H&S
     expected = [[
-        np.array([2.19444444, 2.19230769, np.nan, 2.26666667, 1.88990826, 1.76363636, 3.85]),
-        np.array([3.87368421, 2.51767677, 3.38429752, 3.66666667, 4.13235294]),
+        np.array([2.19444444, 2.19230769, 2.26666667, 1.88990826, 1.76363636, 3.85]),
+        np.array([3.87368421, 2.51767677, np.nan, 3.38429752, 3.66666667, 4.13235294]),
     ]]
     actual = cube.scale_means(hs_dims=[0, 1])
     assert_scale_means_equal(actual, expected)
@@ -180,8 +180,8 @@ def test_cat_x_cat_with_hs_on_both_dims():
 
     # Test with H&S
     expected =[[
-        np.array([2.19444444, 2.19230769, np.nan, 2.26666667, 1.88990826, 1.76363636, 3.85]),
-        np.array([3.87368421, 2.51767677, 3.38429752, np.nan, 3.66666667, 4.13235294]),
+        np.array([2.19444444, 2.19230769, 2.26666667, np.nan, 1.88990826, 1.76363636, 3.85]),
+        np.array([3.87368421, 2.51767677, np.nan, 3.38429752, 3.66666667, 4.13235294]),
     ]]
     actual = cube.scale_means(hs_dims=[0, 1])
     assert_scale_means_equal(actual, expected)